Background Tumor associated macrophages (TAM) have been found to exert a number of pro-tumorigenic including suppression of immune effectors. These properties have been largely associated with an M2 phenotype, while an M1 phenotype has anti-tumor effects. We therefore screened CAR constructs for effectiveness at reverse polarizing M2 phenotype macrophages toward the M1 phenotype.

Methods Over 150 CAR signaling domains, prevalent in monocytes and macrophages and other innate immune effectors, were derived from published sequences. Constructs were designed with an scFv for human mesothelin and hinge domain derived from human CD34. Initial functional screening was performed in HEK-MD2 and HEK-dectin reporter cell lines transfected with DNA. Constructs were considered to have antigen-specific function if mesothelin triggered greater than 20% response over controls. Functional constructs were then stably transduced using a lentiviral vector into THP-1 cell lines and tested in two assays. One assay was a flow cytometry-based phagocytosis assay using pHrodo^TM labeled 5 µm polystyrene beads coated with mesothelin or HER2 (control) proteins. The second was a reverse polarization assay in which stably transduced THP-1 cell lines were first differentiated with PMA and then polarized with IL-4 to an M2 phenotype. The M2 polarized cells were then stimulated with beads coated with mesothelin or HER2 and qPCR was performed for M1 signature genes involved in anti-tumor effects. Data are expressed as ratio of bead versus no beads.

Results Functional testing was on the following constructs: TLR4, IL-2RΥ, FCεR1, control construct with a deleted signaling domain (ΔICD), and a dual construct containing TLR4 and CD3ζ. Initial data are shown in table 1 (table 1).

Conclusions The FCεR1 construct had the highest phagocytosis but lowest upregulation of M1 genes in M2 polarized macrophages. On the other hand, the TLR4 construct had the highest upregulation of M1 genes but lowest phagocytosis. Interestingly, a dual construct consisting of TLR4 and CD3ζ resulted in enhanced phagocytosis when compared to TLR4 alone. Further validation of these constructs with in vitro cellular based phagocytosis, cytotoxicity, and antigen presentation assays are in process. In vivo experiments with nanoparticle delivery of the constructs in immunocompetent mouse models (+/- PD-1 blockade) to fully explore the ability of these CAR constructs to engage T cell responses to neoantigens are ongoing. The assays presented here will allow us to extend the functional testing to our library of CAR construct designs and ultimately identify the ideal construct to pursue for in human testing.

Ethics Approval All animal experiments were conducted with approval of the Brigham Young University IACUC.